Motor Imagery Based Multimodal Biometric User Authentication System Using EEG

Brain Computer Interfaces (BCIs) are regarded as the potential method that bridges the gap between the human brain and the external world. Non-invasive electroencephalographic (EEG) signals are highly individualistic and hence show potential for effective biometric systems. The presented study analyzed the EEG signals for characteristic features elicited by movement and imagination of 4 different upper limb movements. The same limb movement imagery tasks were compared for performance and its validity in developing an effective multimodal biometric system for individuals with motor disabilities. The study involved 10 subjects executing imagined lifting of left and right hands and clenching left and right-hand fists. Along with imagined movement (Motor Imagery), data for actual limb movement was collected, and the performance was compared for both imaginary and actual movement. The proposed pipeline achieved less than 2% False acceptance rate for all the imaginary and actual actions. A novel multimodal approach combining different Motor Imagery (MI) actions was successfully implemented with 98.28% accuracy. Moreover, both imaginary and actual movements showed equally good capability for biometrics purposes suggesting the usability of the proposed biometrics system for people who lost motor abilities or people with poor motor imagery skills.